Clinical experience with immunomodulators in patients with sepsis has been disappointing to date. Tumor necrosis factor soluble receptor (TNFsr) is one such agent. In early preclinical work utilizing endotoxin or gram-negative bacteria infusion challenges, TNFsr was protective. However, in clinical sepsis trials this agent appeared harmful. These findings suggest that factors not yet identified may alter the effects of immunomodulation with such agents. Such factors may include the site, severity, and type of infecting bacteria. In initial studies in a rat model, we found that site and severity of infection did not appear to alter beneficial effects of TNFsr with E. coli infection. In a subsequent study, we looked at the effects of TNFsr with S. aureus. Our results show that across differing doses and sites of bacterial challenge, TNFsr is less beneficial, and sometimes harmful with S. aureus compared to E. coli. We are at present conducting studies to determine why TNFsr might have differing effects in E. coli and S. aureus infections.